Log files are usually internally built within a buffer (possibly ring buffer in the case of syslog), and are flushed to disk when an internal limit (be that size, time, system load or combination of resource factors) is reached and the file is then flushed (written) to disk.You could possibly try:  A> Check the log file you require is not too big or hasn't some corrupted (i.e.: non printable) text, causing it to take more than usual time to overcome this impediment in loading the file for viewing.  One possible test is to start with a log file of just a few lines and see how long that takes to load. If it still takes too long then you may be able to set the logging daemon to some parameter which causes it to force-ably flush to disk more often in line with your requirements.  Check out the logging daemon man file doc. Also use the 'time' command to measure the actual time it takes to respond, which ensures' your got a objective time measure that you can judge with each modification rather than some subjective emotive measure.   B> You could arbitrarily force a "sync" which would flush the buffer to disk, within your code itself. This is what kernel debugging print statements do.  C> Look at your own code, to see if your couldn't add the logging functionality you require directly under your control.